Race, Skin Color, and Common Ancestry

"two key facts can be true at the same time: an apportionment of variance that implies the differentiation among artificially labelled groups is minuscule at any typical genetic locus does not preclude using all the loci together to identify groupings of individuals." — John Novembre, 2022¹

"Virtually all evolutionists would now say that the various people groups did not have separate origins." — The Creation Answers Book, 2019²

Summary

Humans are one species, and a genetically shallow one. Most of our variation lies inside every population: 85.4 percent in Richard Lewontin's 1972 survey of protein and blood-group genes, 84.4 percent in the first repeat with DNA in 1997, and 93 to 95 percent in Noah Rosenberg's 2002 survey of 1,056 people.³⁴⁵ Humans are among the least diverse of the great apes: 79 ape genomes from several species carry more than twice the variable sites found in more than a thousand humans.⁶

That does not make ancestry invisible. A. W. F. Edwards was right in 2003 that many small differences add up. A clustering program given no labels sorts people into groups that match continents, and with many thousands of markers two people from different continents are "never" more alike than two people from the same population.⁵⁷⁸ Both facts are true at once, and they do not contradict each other.¹ The clusters come from small differences in how common each variant is, with people of mixed membership between them, and the traits people notice, such as skin color, are the exceptions rather than a sample of the genome.³⁵⁹ Whether to call the clusters "races" is a choice of words. Nothing in the data ranks them.

The single-origin question splits in two. Separate origins for the races, the polygenism of Louis Agassiz and Josiah Nott, is dead, and every side now agrees. A single origin as a population is what the genetics shows, and both the mainstream recent-African-origin model and the young-earth one-family reading predict the shallow, nested pattern above. The pattern does not choose between them. What chooses is clock-dependent data: allele ages of hundreds of thousands of years, the geography of diversity, and radiocarbon-dated ancient genomes. On those the mainstream reading fits better today, and the young-earth side has no published answer to several of them.

The two articles under this heading:

The Genetics of Skin Color. Skin color is controlled by many genes, not the two of the creationist Punnett square. Populations can change color fast on variation they already carry, but one mid-brown couple cannot produce the human range in one generation, and the light and dark alleles are dated to hundreds of thousands of years.
Scientific Racism and Monogenesis. The polygenists were wrong and the monogenists, mostly Christians reading Genesis, were right. On conduct, nobody comes out clean: Darwin, Haeckel and Galton ranked races, and many Christians, including early creationists, used a misreading of Genesis 9 to do the same, before the movement corrected itself from the text.

Related articles elsewhere: whether one couple fits the genetic data is tested in Adam, Eve, and Human Genetic Diversity, and where the human dispersal began in Out of Africa or Out of Babel?, both under Human Genetic Ancestry. The moral history continues in Christianity, Slavery, and Human Rights and, for one church, in Race, Priesthood, and Changing Revelation.

Part 1: How much humans vary, and whether the variation clusters

Two measurements answer this question, and they look as if they disagree. One asks how the variation at a single gene divides between people and groups. The other asks whether all the genes together can tell groups apart. The first says groups differ very little. The second says they can still be told apart. Both are right.

Most of the variation is inside every population

In 1972 the population geneticist Richard Lewontin took 17 protein and blood-group genes, grouped the world's populations into seven conventional races, and asked how the total diversity at each gene divides. He found that the "proportion of the total species diversity that is contained within populations is 85.4%, with a maximum of 99.7% for the Xm gene, and a minimum of 63.6% for Duffy. Less than 15% of all human genetic diversity is accounted for by differences between human groups!" Differences between populations inside a race added 8.3 percent, "so that only 6.3% is accounted for by racial classification."³ He warned that his method gave too much weight to small, isolated populations, and so overstated the share between groups.³ Rasmus Winther redid every one of Lewontin's calculations in 2022 and got 86, 7 and 7 percent.¹⁰ The errors he found were small and not systematic, and correcting them moves the share between races up from 6.3 to about 7 percent.

DNA gave the same answer. Guido Barbujani and colleagues measured 109 DNA markers in 16 populations in 1997 and found that "differences between members of the same population account for 84.4% of the total". Only one of the 109 markers had less than half its variation within populations.⁴ Rosenberg's team genotyped 377 markers in 1,056 people from 52 populations in 2002: "Within-population differences among individuals account for 93 to 95% of genetic variation; differences among major groups constitute only 3 to 5%."⁵ Of the 4,199 variants they found, only 7.4 percent occurred in one region alone, and those were usually rare even there.⁵ The share depends on the markers and the populations sampled, but every study puts the large majority inside populations.

John Novembre of the University of Chicago put the result in plain terms in a review for its fiftieth anniversary. Lewontin's split "implies that two random individuals from a given ancestry will have on average 15% fewer pairwise differences than individuals sampled across the globe."¹ Two people of the same ancestry differ, on average, about 85 percent as much as two people drawn from anywhere in the world.

Humans also have little variation for a great ape. The 2013 great-ape survey reported that "sequencing of 79 great ape genomes identifies more than double the number of SNPs obtained from the recent sequencing of more than a thousand diverse humans".⁶

What the young-earth books say about it

The young-earth literature reports this result accurately and leans on it. The Creation Answers Book, published by Creation Ministries International, says: "The DNA of any two people in the world typically differs by just 0.2%. Of this, only 6% (i.e. a minuscule 0.012%) can be linked to 'racial' categories; the rest is 'within race' variation."² Its figure on the next page divides the 0.2 percent into 85, 9 and 6 percent, which is Lewontin's 85.4, 8.3 and 6.3 rounded, though the book sources it to a 1994 Discover article rather than to Lewontin.²

The 0.2 percent holds up. The 1000 Genomes Project found that "a typical genome differs from the reference human genome at 4.1 million to 5.0 million sites". Against the book's own figure of 3 billion base pairs, that is 0.14 to 0.17 percent. The same paragraph adds structural variants affecting about 20 million more bases, which would raise the share of letters that differ to about 0.8 percent.¹¹¹² Either way the difference is small.

The 0.012 percent is fair as an average and misleading as a picture. Read the way Novembre reads Lewontin, it says that people from different races differ at somewhat more sites than people from the same population, and that the race grouping accounts for about 6 percent of the difference between two people. That is a reasonable use of Lewontin's numbers. What it cannot mean is that some particular 0.012 percent of the genome is "racial" and the rest is not. Almost no variant belongs to one continent alone.⁵ The between-group share is spread thinly over nearly every variable site, as small differences in how common each variant is. The book draws the right conclusion from it, that the differences are small. What the figure does not show is the next point.

Many genes together do sort people by continent

In 2003 the Cambridge statistician A. W. F. Edwards called Lewontin's inference a fallacy. The 85 percent is real, he said, but the conclusion drawn from it "is unwarranted because the argument ignores the fact that most of the information that distinguishes populations is hidden in the correlation structure of the data and not simply in the variation of the individual factors."⁷ Each small difference is nearly useless alone, but many of them point the same way. Novembre's verdict on Edwards and his predecessors: "They were correct that with large numbers of loci, assignment to groups is possible even if the groups are only subtly distinguished".¹

The data bear Edwards out. Rosenberg's 2002 program, given no information about where anyone came from, "identified six main genetic clusters, five of which correspond to major geographic regions".⁵ The sixth was not a continent. It "consisted largely of individuals of the isolated Kalash group", a small Indo-European-speaking people in northwest Pakistan.⁵ D. J. Witherspoon's team measured how often two people from different populations are more alike than two from the same one. With 10 markers the answer is "nearly one-third of the time", with 100 about 20 percent, and with 1,000 about 10 percent.⁸ "However, if genetic similarity is measured over many thousands of loci, the answer becomes 'never' when individuals are sampled from geographically separated populations."⁸

Lewontin never denied that groups exist. The first page of his 1972 paper says that "variation is nodal, that individuals fall in clusters in the space of phenotypic description, and that those clusters, which we call demes, or races, or species, are the outcome of an evolutionary process acting on the individual variation."³ A decade later he wrote that "No one would mistake a Chinese for a West African or a Finn for an Australian aborigine".¹ His claim was about how much the race label predicts at a typical gene, not about whether ancestry can be detected.¹

What the two facts mean together

Novembre's summary: an apportionment that makes group differences "minuscule at any typical genetic locus does not preclude using all the loci together to identify groupings of individuals." Lewontin and Edwards were "addressing two closely related problems, neither actually contradicting the other."¹¹

Three further results say what kind of groups these are.

They are gradations, not types. Rosenberg's team wrote that "genetic differences among human populations derive mainly from gradations in allele frequencies rather than from distinctive 'diagnostic' genotypes", and that structure appeared "only in the accumulation of small allele-frequency differences across many loci".⁵
Their edges are steps at barriers, with mixed people between. A 2005 follow-up with 993 markers found the clusters arise "from small discontinuous jumps in genetic distance for most population pairs on opposite sides of geographic barriers", with "some individuals from intermediate geographic locations having mixed membership".⁹
The traits people notice are the outliers. Lewontin argued that racial classification "gives tremendous emphasis to those characters to which human perceptions are most finely tuned (nose, lip and eye shapes, skin color, hair form and quantity)".³ His question was "are these large differences typical?"¹ At skin-color genes they are not. The light allele at SLC45A2 makes up 97 percent of the copies carried by Europeans today and is absent from one study's African reference sample.¹³¹⁴ That is far beyond the 6 percent average. Witherspoon's team drew the corollary for traits controlled by only a few genes: "Phenotypes controlled by a dozen or fewer loci can therefore be expected to show substantial overlap between human populations."⁸

So the honest answer has two halves. The continental ancestry of people from distant regions can be read from DNA, reliably, with enough markers. And the groups so read differ a little, by gradual steps, with most of each person's variation shared with people everywhere. Neither half ranks anyone, and neither half supports the nineteenth-century idea of races as deep, separately originated types.

Part 2: The genetics of skin color

The question. The popular creationist books say that two mid-brown parents, heterozygous at a few genes, can produce the whole human range of skin color in one generation, so that a mid-brown first couple could be the ancestors of every color. The Creation Answers Book works it with two genes: "For simplicity, let's assume there are only two".² One Blood concludes from its Punnett square that "an entire range of 'colors,' from very white to very black, can result in only one generation".¹⁵ Is that how skin color works, and how fast can it change?

The verdict. Skin color is controlled by many genes, not two. Four regions explain about 29 percent of the inherited variation among Africans, and in the KhoeSan, where heritability is 0.97, the known pigment genes explain almost none of it and "more than 50 loci" are implicated.¹⁶¹⁷ Populations can change color fast on variation they already carry: on the Pontic-Caspian steppe the light SLC45A2 allele rose from 0.432 to 0.927 in about 5,000 years.¹⁴ But one couple cannot produce the range in one generation. The child article's own arithmetic, using measured effect sizes, finds that even eight genes put 73 percent of such a couple's children in the middle of the range, and that adding genes makes the extremes rarer, not commoner, the reverse of what the book's footnote 9 claims.² The hard problem is age, not speed: light and dark alleles at MFSD12 and OCA2 are dated to between about 600,000 and 1.7 million years, and a Spanish hunter dated to 7,940 to 7,690 calibrated radiocarbon years ago lacked both main European light alleles.¹³¹³¹⁶¹⁶

Where the young-earth argument now stands. Creation Ministries International has already moved. Andrew Sibley wrote in 2025 that "it makes sense to accept that the early European immigrants most likely had darker skin than those of today" and placed the disagreement on "the inflated timescale."¹⁸¹⁸ That is a dispute about dating, not about pigment genetics. Read the full argument in The Genetics of Skin Color.

Part 3: Scientific racism and monogenesis

The question. Who was right about whether all people share one origin, and who used science or Scripture to rank the races?

The verdict. On origins, the monogenists were right. The American polygenists of the 1840s and 1850s held that the races were created separately: "men must have originated in nations, as the bees have originated in swarms", in Louis Agassiz's words.¹⁹ Darwin predicted the dispute "will die a silent and unobserved death", and it did.²⁰ On conduct, both sides failed. The same Darwin expected "the civilised races of man" to "exterminate and replace throughout the world the savage races".²⁰ The slave South mostly refused polygenism as "an assault upon orthodox religion", and then defended slavery from the curse on Ham.²¹ Early creationists carried the racial ideas of their day: Henry Morris in 1976 assigned Hamites, "especially the Negroes", "a genetic character concerned mainly with mundane, practical matters".²² The correction came from the text. One Blood (1999) says that "The curse of Canaan has nothing whatsoever to do with skin color".²³ The record supports neither the claim that evolution invented racism nor the claim that the Bible did. Read the full record in Scientific Racism and Monogenesis.

Part 4: What this adds up to for a single origin

The two readings, in their own words

The mainstream reading is a recent African origin. Modern humans arose in Africa, Crawford's team puts that origin at about 300,000 years ago, and a subset of Africans later spread across the rest of the world.¹⁶ The young-earth reading is one family. The Creation Answers Book says that all humans descend from Noah's family, that God broke the population into groups at Babel, and that "variations in skin colour, for example, can develop in only a few generations. There is good evidence that the various people groups we have today have not been separated for huge periods of time."² CMI's Andrew Sibley dates the first settlers of Britain to "2000–2400 BC".¹⁸

The same book states the common ground fairly. Evolutionists "would agree with the biblical creationist that all people groups have come from the same original population. Of course, they say that such groups as the Aborigines and the Chinese have had many tens of thousands of years of separation."² That is the disagreement in one sentence: one origin is agreed, and the clock is not.

What both readings predict

Most of what this heading's articles found is predicted by both readings. A population that began small and spread recently, whether 4,500 or 300,000 years ago, should show little variation, most of it shared, with the newer and rarer variants confined to the regions where they arose.

Observation	Recent African origin	Young-earth one family	What the data show
One interbreeding species, no separate origins	Predicted	Predicted	Polygenism refuted; agreed by both sides
Most variation inside every population	Predicted	Predicted	84 to 95 percent within
Less variation than the great apes	Predicted, from a small long-term population	Predicted, from few founders	Humans among the least diverse great apes
Common variants shared everywhere, rare ones local	Predicted	Predicted	Found by the 1000 Genomes Project
Visible traits able to shift fast	Allowed	Required	Measured: 0.432 to 0.927 at SLC45A2 in about 5,000 years
Africa the most diverse, others a subset	Predicted	Must be explained	Found
Allele ages of hundreds of thousands of years	Predicted	Must be explained	Found at skin-color genes and genome-wide
Archaic admixture about 2,000 generations ago	Predicted	Must be explained	Found

The 1000 Genomes Project reported both of the nested facts in one paragraph: "Individuals from African ancestry populations harbour the greatest numbers of variant sites", and "Although most common variants are shared across the world, rarer variants are typically restricted to closely related populations".¹¹

The amount of difference between populations does not need deep time either. Genetic drift alone, in isolated groups, builds between-group differences quickly when the groups are small. To reach Barbujani's 15.6 percent between populations in the 130 to 200 generations since a Babel dispersal 4,000 to 5,000 years ago, isolated groups would need about 380 to 590 breeding adults each. On the mainstream's roughly 2,000 generations, the same arithmetic needs groups of about 5,900.²⁴ Both sizes are possible. So the level of differentiation, like the shallowness, fits either reading. This is a sketch: it ignores migration, which would require smaller groups, and founder events, which would allow larger ones.

Where the two readings part

The readings make different predictions only where a clock or a map enters, and those are tested in other articles.

Depth. The light and dark skin-color alleles at MFSD12 and OCA2 coalesce between about 600,000 and 1.7 million years ago on the standard clock.¹⁶¹⁶ Genome-wide, Adam, Eve, and Human Genetic Diversity reports a median of 495,000 years back to four surviving lineages. The young-earth answer is created diversity, which can bridge any gap by definition, and no published calculation on that side has met these ages. Crawford's own paper names balancing selection as the explanation for the deep OCA2 lineages, which is an opening a young-earth critic could test, and nobody has.
Place. Out of Africa or Out of Babel? refits the decline of diversity with distance and finds the source in Africa, not Mesopotamia, with an R-squared of 0.859 from Addis Ababa against 0.647 from Babylon. It also finds that the non-African populations alone cannot tell the two apart. A Near Eastern staging post for the non-African world fits the data on both readings. The same article's hardest number is a clock that needs no calendar: 1,805 to 2,043 generations of recombination since Neanderthal admixture, against about 160 generations in 4,000 years.
Founders. One couple against a long-term effective population of about ten thousand. The same Adam and Eve article finds that three of the four standard arguments against a single couple do not test it, and that two, the HLA-DRB1 lineage count and the 495,000-year median, have no young-earth answer in print.
Dated ancient genomes. La Braña 1, at 7,940 to 7,690 calibrated radiocarbon years before present, lacked both main European light alleles.¹³¹³ On a biblical chronology anyone living in Europe lived after Babel, so his radiocarbon age has to fit inside a few thousand calendar years. The genotype is a direct call. What makes it a problem is the radiocarbon date, which is exactly what Sibley disputes.

So the race question and the age question come apart. The genetics settles the race question: one origin, one shallow population, no separately created types. Genesis and Acts were right about that, and the mainstream now says the same. The age question is not settled by the shallowness, and on the clock-dependent evidence the young-earth side carries the burden. That evidence belongs to the Human Genetic Ancestry articles and to the dating methods, not to the study of race.

Objections and responses

"Lewontin proved that race has no biological basis."

He proved that a race label predicts little at a typical gene, and his numbers have held for fifty years.³¹⁰ He did not prove that ancestry cannot be read from DNA, and Novembre's review finds he did not claim to.¹ His own paper says variation is "nodal" and falls into clusters.³ With thousands of markers, two people from distant populations are never more alike than two from the same one.⁸ "No biological basis" overstates his result. "No basis for ranking" states it correctly.

"Edwards proved the races are real, so the old racial taxonomy was right."

Edwards proved that groups which overlap at every gene can still be told apart when many genes are combined.⁷ The groups so found are gradations in frequency with mixed people between them, and the number of clusters is the analyst's choice: Rosenberg's sixth cluster was one small Pakistani people, not a continent.⁵⁵ The old taxonomy claimed fixed types with inherited characters of mind and temperament. Nothing in Edwards's argument, or in any clustering study, supports that. The polygenist version, separate creation, is refuted outright.

"Race is only 0.012 percent of our DNA, so it means nothing biologically."

The figure is The Creation Answers Book's, and its underlying numbers are Lewontin's, reported accurately.²² As an average it is fair: the race grouping accounts for about 6 percent of the difference between two people. The book uses it only to say that the differences are small, which is true. The skeptic's further step does not follow. A small average share, spread over millions of sites, still identifies continental ancestry when the sites are read together, as Rosenberg and Edwards showed.⁵⁷

"Low human diversity proves a recent creation."

It fits a recent creation. It also fits a population that stayed small, about ten thousand breeders, for a long time, and the mainstream explains it that way. Shallowness alone does not tell the two apart, and neither does the level of between-group difference.²⁴ The clocks do, and they are argued in the articles linked in Part 4.

"Shallow, nested variation proves Out of Africa and refutes Genesis."

The shallowness and the nesting are predicted by both readings. What favors the African reading is the position of African diversity, the allele ages and the admixture generation count, all of which depend on calibrated mutation and recombination clocks. The young-earth side can dispute those clocks. It cannot dispute the pattern, and it does not try to: The Creation Answers Book uses the same numbers.²

"If skin color can change in a few thousand years, the mainstream is wrong about time."

Speed is not where the timescales conflict. The mainstream measured the fast change itself, on the steppe and in Britain.¹⁴ The conflict is the age of the alleles and the radiocarbon dates of ancient genomes that lack them.¹³¹⁶

"Early creationists were racists, so the young-earth position is tainted."

Several were, and the child article quotes them.²² So were Darwin and many of the evolutionists who followed him.²⁰ The conduct of a view's holders does not settle whether the view is true. The genetic facts in Part 1 are the same whoever measured them, and the creationist movement's correction came from its own text.²³

"One origin as a population is not Adam and Eve."

Correct. The genetics shows one population, not one couple. Whether a single couple fits is the question of Adam, Eve, and Human Genetic Diversity, and this hub claims only what the race data show.

Issues and unanswered questions

Edwards 2003 was read only in its abstract. The full text is paywalled and no open copy was found. His argument is quoted from the abstract and described through Novembre's 2022 review. The abstract states the whole claim, but the full paper would let his worked example be quoted.
Lewontin 1972 was read through archive.org's search-inside text. The page numbers rest on running heads read at three places in the chapter, and they agree with Novembre's citation of the final sentence to page 397.
The drift arithmetic in Part 4 is a sketch. It assumes isolated groups of constant size with no migration and no selection. Nobody on either side has fitted a Babel dispersal, with realistic migration and founder events, to the measured differentiation between human populations. Out of Africa or Out of Babel? refits the closest published measurement, the decline of diversity with distance.
How much more differentiated are visible-trait genes than the average? Part 1 shows it for one locus from two studies. A genome-wide comparison of pigmentation genes against the average between-group share was not assembled here.
Coalescent ages under balancing selection. Would a standard coalescent estimator return 600,000 years at a locus under strong balancing selection in a population founded a few thousand years ago? Nobody on either side has run this.
What a cluster means is still argued. Whether to call continental clusters "races", "ancestry groups" or "populations" is a question of definition that genetics does not settle. This hub uses "race" only for the historical categories.
No contradictions between the two children were found. Each states the genetics of one origin consistently with the other: one shallow population, most variation within groups, skin color many-gened, and the young-earth dispute located in chronology.

Verdict

Humans are one shallow species. Eighty-four to ninety-five percent of genetic variation lies within every population, almost no variant belongs to one continent alone, and humans are among the least diverse of the great apes. Lewontin measured this in 1972, it has held for every kind of marker since, and the young-earth literature reports it accurately.

Ancestry is still written in the genome. Edwards was right that many small differences add up. With thousands of markers, two people from distant populations are never more alike than two from the same one, and clustering programs recover continents without being told them. Both results are true at once. The groups they describe are gradations with mixed people at their edges, the visible traits are the exceptions, and nothing in the data ranks anyone.

The nineteenth-century debate is over, and Genesis won it. Separate creation of the races failed as science, as the child on scientific racism shows. Skin color, the trait most people mean by race, is controlled by many genes and can shift fast in a population, though not in one family in one generation.

A shallow, shared human genome is predicted by both modern readings of human origins. A recent African origin and a young-earth family both expect it, and the level of difference between groups fits either. They part on the clocks: the ages of alleles, the generations since archaic admixture, the map of diversity and the radiocarbon dates of ancient genomes. On those the mainstream reading fits better today, and the young-earth side has yet to publish answers to several of them. An honest young-earth position takes the race question as won, drops the two-gene Punnett square, and argues the chronology, which is where Creation Ministries International has already gone.

Sources

1.
Novembre, John. “The background and legacy of Lewontin's apportionment of human genetic diversity.” Philosophical Transactions of the Royal Society B 377(1852):20200406. 2022. A review by a University of Chicago population geneticist for the fiftieth anniversary of Lewontin's 1972 paper. It sets out what Lewontin measured and concluded, how the 85% figure has replicated, and the multi-locus criticisms from Spielman and Smouse in 1976 to Edwards in 2003. Mirror: PMC
1. 1a.
  Page 6: “They were correct that with large numbers of loci, assignment to groups is possible even if the groups are only subtly distinguished (i.e. classification using a genetically informed taxonomy is possible).” "They" are the critics of Lewontin's inference: Spielman and Smouse, Mitton, and Edwards. ↩
2. 1b.
  Page 6: “No one would mistake a Chinese for a West African or a Finn for an Australian aborigine”. “The question is, are these large differences typical?” Novembre quoting Lewontin's 1982 book Human Diversity, pages 111 and 117. ↩ ↩
3. 1c.
  Page 7: “the 85%/15% result of Lewontin implies that two random individuals from a given ancestry will have on average 15% fewer pairwise differences than individuals sampled across the globe.” ↩
4. 1d.
  Page 7: “two key facts can be true at the same time: an apportionment of variance that implies the differentiation among artificially labelled groups is minuscule at any typical genetic locus does not preclude using all the loci together to identify groupings of individuals.” ↩ ↩
5. 1e.
  Page 7: “As such, both Lewontin and his apparent critics like Edwards can be understood as addressing two closely related problems, neither actually contradicting the other.” ↩ ↩ ↩ ↩
2.
Catchpoole, David et al. “Chapter 18: How did all the different 'races' arise (from Noah's family)?” The Creation Answers Book, 223–238. Creation Book Publishers. Edited by Don Batten. 2019. A chapter from a creationist question-and-answer book, edited by Don Batten, arguing that skin colour variation can develop in only a few generations after the people of Noah's family spread out from Babel. Its skin colour argument rests on a simplified two-gene model, and it says natural selection only sorts the colour capacity that was already there. This hub uses it as the clearest current young-earth statement on race, skin colour and the timescale of human diversity.
1. 2a.
  Page 228: “Skin colour is governed by more than one pair of genes. For simplicity, let's assume there are only two, located at positions A and B on the chromosomes.” ↩
2. 2b.
  Page 228: This is footnote 9 of the chapter, and its superscript sits on "only two" in the body text on the same page. “This simplification is not done to help our case—the more genes there are, the easier it is to have a huge range of 'different' colours. The principle involved can be understood by using two as an example.” The first half is right and the second is the load-bearing defence of the two-locus model. Under the chapter's own additive model the probability of an extreme child is (1/4)ⁿ and the offspring standard deviation shrinks as 1/√n relative to the range, so more loci make the extremes rarer. Worked out in The Genetics of Skin Color, Part 5. ↩
3. 2c.
  Pages 223–224: “Modern genetics shows how, following such a break-up of a population, variations in skin colour, for example, can develop in only a few generations. There is good evidence that the various people groups we have today have not been separated for huge periods of time.” ↩
4. 2d.
  Page 224: “The DNA of any two people in the world typically differs by just 0.2%.” “Of this, only 6% (i.e. a minuscule 0.012%) can be linked to ‘racial’ categories; the rest is ‘within race’ variation.” The book's footnote to the first sentence cites J. C. Gutin, "End of the rainbow", Discover, November 1994, pages 71–75. ↩ ↩ ↩
5. 2e.
  Page 224: “Virtually all evolutionists would now say that the various people groups did not have separate origins.” “So they would agree with the biblical creationist that all people groups have come from the same original population. Of course, they say that such groups as the Aborigines and the Chinese have had many tens of thousands of years of separation.” ↩ ↩
6. 2f.
  Page 225, top: A diagram headed Human DNA (3,000,000,000 base pairs) divides the 0.2% variation between individuals into 85% within a local ethnic group (e.g. Cantonese), 9% between ethnic and linguistic groups within a 'race' (e.g. Cantonese and Japanese), and 6% between 'races' (e.g. Asian and Caucasian). Its caption: “The variation in DNA between human individuals shows that racial differences are tiny.” ↩ ↩ ↩
3.
Lewontin, Richard C. “The Apportionment of Human Diversity.” Evolutionary Biology 6:381–398. Edited by Theodosius Dobzhansky, Max K. Hecht and William C. Steere. 1972. The founding study of how human genetic diversity divides. Lewontin, a leading population geneticist and an outspoken critic of racial classification, took gene frequencies for 17 protein and blood-group genes in populations grouped into seven conventional races, and split the diversity at each gene with Shannon's information measure into shares within populations, between populations of a race, and between races. He states that his weighting of small populations overstates the share between groups. Mirror: Archive.org
1. 3a.
  Page 381: “that variation is nodal, that individuals fall in clusters in the space of phenotypic description, and that those clusters, which we call demes, or races, or species, are the outcome of an evolutionary process acting on the individual variation.” ↩ ↩
2. 3b.
  Page 382: “The erection of racial classification in man based upon certain manifest morphological traits gives tremendous emphasis to those characters to which human perceptions are most finely tuned (nose, lip and eye shapes, skin color, hair form and quantity), precisely because they are the characters that men ordinarily use to distinguish individuals.” ↩ ↩
3. 3c.
  Page 389: “These conventions also overestimate the proportion of the total diversity that is between populations and races as opposed to within populations since it gives too much weight to small isolated populations and to less numerous races like the Amerinds and Australian aborigines, both of which have gene frequencies that differ markedly from the rest of the species.” ↩
4. 3d.
  Page 396: “proportion of the total species diversity that is contained within populations is 85.4%, with a maximum of 99.7% for the Xm gene, and a minimum of 63.6% for Duffy. Less than 15% of all human genetic diversity is accounted for by differences between human groups! Moreover, the difference between populations within a race accounts for an additional 8.3%, so that only 6.3% is accounted for by racial classification.” ↩ ↩ ↩
4.
Barbujani, Guido et al. “An apportionment of human DNA diversity.” Proceedings of the National Academy of Sciences 94(9):4516–4519. 1997. The first repeat of Lewontin's analysis with DNA: 109 markers, microsatellites and restriction sites, in 16 populations worldwide. It found the same apportionment as the protein studies. Mirror: PMC
1. 4a.
  Page 4516, section Abstract: “differences between members of the same population account for 84.4% of the total”. Section Discussion: “However, only for one of the 109 loci studied was the within-population component of variance less than 50% of the total.” ↩ ↩ ↩
5.
Rosenberg, Noah A. et al. “Genetic Structure of Human Populations.” Science 298(5602):2381–2385. 2002. A genotyping study of 377 microsatellite markers in 1,056 people from the 52 populations of the Human Genome Diversity Panel. It measured the apportionment of variation and ran a clustering program that grouped individuals without being told where they came from. Mirror: Web.stanford.edu
1. 5a.
  Page 2381, section Abstract: “Within-population differences among individuals account for 93 to 95% of genetic variation; differences among major groups constitute only 3 to 5%. Nevertheless, without using prior information about the origins of individuals, we identified six main genetic clusters, five of which correspond to major geographic regions”. ↩ ↩ ↩ ↩ ↩
2. 5b.
  Page 2382: “Only 7.4% of these 4199 alleles were exclusive to one region; region-specific alleles were usually rare, with a median relative frequency of 1.0% in their region of occurrence”. ↩ ↩
3. 5c.
  Page 2382: “clusters corresponded largely to major geographic regions.” “did not match a major region but consisted largely of individuals of the isolated Kalash group, who speak an Indo-European language and live in northwest Pakistan”. The first sentence describes the run with five clusters, the second the sixth cluster added in the next run. ↩ ↩
4. 5d.
  Page 2384: “Because most alleles are widespread, genetic differences among human populations derive mainly from gradations in allele frequencies rather than from distinctive “diagnostic” genotypes. Indeed, it was only in the accumulation of small allele-frequency differences across many loci that population structure was identified.” ↩ ↩ ↩
6.
Prado-Martinez, Javier et al. “Great ape genetic diversity and population history.” Nature 499:471–475. 2013. Whole-genome sequences of 79 wild and captive chimpanzees, bonobos, gorillas and orangutans, compared with nine human genomes. It found humans among the least genetically diverse of the great apes. Mirror: PMC
1. 6a.
  Section Discussion: “Humans, Western chimpanzees, and Eastern gorillas all show a remarkable dearth of genetic diversity when compared to other great apes. It is striking, for example, that sequencing of 79 great ape genomes identifies more than double the number of SNPs obtained from the recent sequencing of more than a thousand diverse humans”. ↩ ↩
7.
Edwards, A. W. F. “Human genetic diversity: Lewontin's fallacy.” BioEssays 25(8):798–801. 2003. A short critique by a Cambridge statistical geneticist. Using a two-population model, Edwards shows that groups which overlap at every single locus can still be told apart when many loci are combined.
1. 7a.
  Page 798, section Abstract: “This conclusion, due to R.C. Lewontin in 1972, is unwarranted because the argument ignores the fact that most of the information that distinguishes populations is hidden in the correlation structure of the data and not simply in the variation of the individual factors.” ↩ ↩ ↩ ↩
8.
Witherspoon, D. J. et al. “Genetic similarities within and between human populations.” Genetics 176(1):351–359. 2007. A study using three data sets of up to several thousand markers to reconcile Lewontin's and Edwards's findings. It measures how often two people from different populations are more alike than two from the same one, as the number of loci grows.
1. 8a.
  Section Discussion: “However, if genetic similarity is measured over many thousands of loci, the answer becomes “never” when individuals are sampled from geographically separated populations.” The same paragraph gives the answer as nearly one-third of the time with 10 loci, about 20% with 100 loci, and about 10% with 1,000 loci. ↩ ↩ ↩
2. 8b.
  Section Discussion: “Given 10 loci, three distinct populations, and the full spectrum of polymorphisms (Figure 2E), the answer is ≅ 0.3, or nearly one-third of the time. With 100 loci, the answer is ∼20% of the time and even using 1000 loci, ≅ 10%.” The question being answered is how often a pair of people from one population is more different than two people from different populations. ↩
3. 8c.
  Section Abstract: “Phenotypes controlled by a dozen or fewer loci can therefore be expected to show substantial overlap between human populations.” ↩
9.
Rosenberg, Noah A. et al. “Clines, clusters, and the effect of study design on the inference of human population structure.” PLoS Genetics 1(6):e70. 2005. A follow-up to the 2002 study with 993 markers, testing whether the continental clusters were an artifact of uneven sampling along gradual clines.
1. 9a.
  Page 0660, section Abstract: “the clusters arise not as an artifact of the sampling scheme, but from small discontinuous jumps in genetic distance for most population pairs on opposite sides of geographic barriers”. “with some individuals from intermediate geographic locations having mixed membership in the clusters that correspond to neighboring regions.” ↩ ↩
10.
Winther, Rasmus Grønfeldt. “Lewontin (1972).” Remapping Race in a Global Context. Routledge, London. penultimate version, arXiv 2110.10945 edition. Edited by Ludovica Lorusso and Rasmus Grønfeldt Winther. 2022. A chapter-length guide to Lewontin's 1972 paper by a philosopher of science. Winther traces the works Lewontin drew on, reads each section closely, and redoes all of Lewontin's diversity calculations from his published data, finding small errors at every gene but one and no systematic bias.
1. 10a.
  Page 3: “In redoing all his calculations, I find that Lewontin’s Distribution of (1) within populations, (2) among populations but within races, and (3) among races diversity apportionments are, respectively and rounding to the nearest percentage, in fact 86%/7%/7%.” ↩ ↩
11.
The 1000 Genomes Project Consortium. “A global reference for human genetic variation.” Nature 526:68–74. 2015. The final report of the 1000 Genomes Project: 2,504 genomes from 26 populations on five continents, the reference survey of human genetic variation. Mirror: PMC
1. 11a.
  Page 69: “Individuals from African ancestry populations harbour the greatest numbers of variant sites, as predicted by the out-of-Africa model of human origins.” “Although most common variants are shared across the world, rarer variants are typically restricted to closely related populations”. ↩
2. 11b.
  Page 68: “We find that a typical genome differs from the reference human genome at 4.1 million to 5.0 million sites”. The same paragraph gives the typical genome 2,100 to 2,500 structural variants, affecting about 20 million bases of sequence. ↩ ↩
12.
Computed here from the 1000 Genomes figures and the book's own genome size of 3 billion base pairs.¹¹² 4.1 million ÷ 3 billion = 0.14 percent, and 5.0 million ÷ 3 billion = 0.17 percent. Adding the roughly 20 million bases affected by structural variants gives 24.1 to 25.0 million ÷ 3 billion = 0.80 to 0.83 percent. These figures compare each genome with the reference genome rather than with another randomly chosen person, but the two comparisons are of the same order. ↩
13.
Olalde, Iñigo et al. “Derived immune and ancestral pigmentation alleles in a 7,000-year-old Mesolithic European.” Nature 507(7491):225–228. Mar 13, 2014. A genome study of a male skeleton from La Braña in León, Spain: the tooth gave 3.40x effective genomic depth of coverage (each part of the genome read about 3.4 times on average) and 48.4% human DNA content. It found that he carried two copies of the ancestral allele (the older version of a gene) at SLC45A2 and SLC24A5, the two strongest known European light-skin loci, and the authors' skin-colour prediction is inferred from his genes rather than measured. The article uses it as the sharpest case of an ancient European without the light alleles. Mirror: PMC
1. 13a.
  Section Main text, paragraph 2: “The skeletons were dated to ~7,000 years BP (7,940–7,690 calibrated BP).” ↩ ↩ ↩
2. 13b.
  Section Main text, paragraph 7: “Of the ten variants, the Mesolithic genome carried the ancestral and non-selected allele as a homozygote in three regions: C12orf29 (a gene with unknown function), SLC45A2 (rs16891982) and SLC24A5 (rs1426654). The latter two variants are the two strongest known loci affecting light skin pigmentation in Europeans and their ancestral alleles and associated haplotypes are either absent or segregate at very low frequencies in extant Europeans (3% and 0% for SLC45A2 and SLC24A5 respectively).” ↩ ↩ ↩
14.
Wilde, Sandra et al. “Direct evidence for positive selection of skin, hair, and eye pigmentation in Europeans during the last 5,000 y.” PNAS 111(13):4832–4837. Apr 1, 2014. An ancient DNA study of 63 of 150 samples from the Eneolithic and Bronze Age Pontic–Caspian steppe, mostly in modern Ukraine, with pigmentation genotypes from 48 of them. The authors compared allele frequencies in these ancient people with those in modern Ukrainians and fitted how strongly selection must have acted to produce the change, rejecting neutrality for every allele studied, with selection of around 2–10% per generation. The article uses it as the direct measurement that these pigmentation alleles were under selection in Europe.
1. 14a.
  Table 1: Derived allele frequencies with equal-tailed 95% intervals, and the sample size (2N, the number of gene copies sampled) in square brackets. HERC2 rs12913832 A>G: ancient 0.160 (0.099–0.247) [94], modern Ukrainian 0.651 (0.546–0.744) [86]. SLC45A2 rs16891982 C>G: ancient 0.432 (0.296–0.578) [44], modern Ukrainian 0.927 (0.849–0.965) [82]. TYR rs1042602 C>A: ancient 0.043 (0.018–0.106) [92], modern Ukrainian 0.367 (0.279–0.466) [98]. All three alleles are at zero in the paper's African reference sample and at 0.002–0.007 in its Asian one. ↩ ↩ ↩
15.
Ham, Ken et al. “Chapter 4: One race.” One Blood: The Biblical Answer to Racism (Master Books, 1999). Feb 8, 2006. A chapter from the book One Blood, by Ken Ham, Don Batten and Carl Wieland, published by Master Books in 1999 and republished on creation.com on 8 February 2006. It argues that all peoples come from one original population, and that two mid-brown parents, each carrying two gene pairs for skin darkness, could have children of every shade in one generation, using a Punnett square (a grid of the gene combinations parents can pass on). The article quotes it for the Punnett-square result, after quoting The Creation Answers Book's version of the same two-gene argument.
1. 15a.
  Immediately after the Punnett square of two mid-brown parents: “Surprisingly, we find that an entire range of 'colors,' from very white to very black, can result in only one generation, beginning with this particular type of mid-brown parents.” The argument's logical form matters: it does not derive the mid-brown ancestors from genetics. It assumes that all the factors were present in Adam and Eve and infers the phenotype, the visible skin colour, from that. ↩
16.
Crawford, Nicholas G. et al. “Loci associated with skin pigmentation identified in African populations.” Science 358(6365):eaan8433. Nov 17, 2017. A genome study of skin colour in Africa, published in Science. The team measured skin reflectance on the inner arm with a colour meter, as a proxy for melanin, in 2,092 people living in Ethiopia, Tanzania and Botswana, and genotyped 1,570 of them on an Illumina Omni5 gene chip. Four gene regions explain about 29% of the inherited variation in skin colour among these Africans, which leaves about 71% elsewhere. The article uses it for the age of the light-skin alleles and for that four-region figure.
1. 16a.
  Section Discussion, first paragraph: “Most alleles associated with light and dark pigmentation in our dataset are estimated to have originated prior to the origin of modern humans ~300 ky ago (26). In contrast to the lack of variation at MC1R, which is under purifying selection in Africa (61), our results indicate that both light and dark alleles at MFSD12, DDB1, OCA2, and HERC2 have been segregating in the hominin lineage for hundreds of thousands of years (Fig. 4). Further, the ancestral allele is associated with light pigmentation in approximately half of the predicted causal SNPs; Neanderthal and Denisovan genome sequences, which diverged from modern human sequences 804 kya (62), contain the ancestral allele at all loci.” “These observations are consistent with the hypothesis that darker pigmentation is a derived trait that originated in the genus Homo within the past ~2 million years after human ancestors lost most of their protective body hair, though these ancestral hominins may have been moderately, rather than darkly, pigmented.” ↩
2. 16b.
  Section Variance explained by the top eight candidate SNPs: “Approximately 28.9% (S.E. 10.6%) of the pigmentation variance is attributable to these SNPs. Considering each locus in turn and all significantly associated variants (p-value < 5 × 10−8), the trait variation attributable to each locus is: SLC24A5 (12.8%, S.E. 3.5%), MFSD12 (4.5%, S.E. 2.1%), DDB1/TMEM138 (2.2%, S.E. 1.5%), and OCA2/HERC2 (3.9%, S.E. 2.9%). Thus ~29% of the additive heritability of skin pigmentation in Africans is due to variation at these four regions. This observation indicates that the genetic architecture of skin pigmentation is simpler (i.e., fewer genes of stronger effect) than other complex traits, such as height.” The phrase creationist writing quotes is "simpler than height". The sentence before it leaves 71% of the heritability elsewhere. ↩
3. 16c.
  Section MFSD12: “Coalescent analysis of the SGDP dataset indicates that the rs10424065 (T) allele predates the 300 kya origin of modern humans (26) (estimated TMRCA of 612 kya, 95% CI 515-736 kya).” “Coalescent analysis of the SGDP dataset indicates that the TMRCA for the derived rs6510760 (A) allele is 996 kya (95% CI 0.82-1.2 mya).” ↩ ↩ ↩
4. 16d.
  Section OCA2/HERC2: On rs1800404. “Coalescent analysis indicates that the TMRCA of all lineages is 1.7 mya (95% CI: 1.5-2.0 mya) and the TMRCA of lineages containing the derived (T) allele is 629 kya (95% CI 426-848 kya). … The deep coalescence of lineages, and the positive Tajima's D values in this region in both African and non-African populations, is consistent with balancing selection acting at this locus.” The interior ellipsis drops two parenthetical figure references. The balancing-selection clause is the paper's own explanation for the depth, and it is also the opening a young-earth critic would need. ↩ ↩
17.
Martin, Alicia R. et al. “An Unexpectedly Complex Architecture for Skin Pigmentation in Africans.” Cell 171(6):1340–1353.e14. Nov 30, 2017. A genetic study of skin colour, in which 465 KhoeSan were phenotyped with a portable reflectance spectrophotometer and genotyped, 278 ǂKhomani San and 187 Nama, alongside a compiled global set of 4,712 individuals from 32 populations. It found that heritability, the share of colour differences that come from genes, was 0.96 to 0.97, while the known pigment genes explained almost none of it, so colour is shaped by many genes with small effects. This hub uses it to show that the simple few-gene model fails in Africa.
1. 17a.
  Section Discussion: “Most of the pigmentation variability in KhoeSan populations is not explained by previously identified loci, suggesting that more than 50 loci (and indeed, likely far more, given our genomic heritability estimates) with a distribution of mostly small effects contribute to variation in pigmentation in the KhoeSan. This suggests that skin pigmentation is a far more complex trait than previously discussed, analogous to numerous other complex traits discussed in biomedical literature.” ↩
18.
Sibley, Andrew. “Were the first Britons dark-skinned?” Creation Ministries International. Mar 24, 2025. A Creation Ministries International article, written in March 2025 on press reports that the early Britons who built Stonehenge had dark skin. It concedes that early European immigrants most likely had darker skin and that lighter alleles spread under weaker UV-B radiation, and it rests the disagreement on the timescale and on radiocarbon dating. Unlike Carter 2017, it cites the primary preprint, Perretti et al. 2025, with the correct DOI, also cites Brace et al. 2019, and reports the preprint's finding accurately; the article uses it for where the creationist argument now stands.
1. 18a.
  Section A biblical creation perspective: “Biblically, it makes sense to accept that the early European immigrants most likely had darker skin than those of today.” “We would agree that the alleles for lighter skin tones probably became dominant because of weaker UV-B radiation at northern latitudes. This is merely to recognise that changes in genes can and do occur.” ↩
2. 18b.
  Section A biblical creation perspective: “Where we especially differ from the secular narrative is over the inflated timescale. Based on the Genesis record, we believe that the first settlers to Britain arrived 2000–2400 BC.” ↩ ↩
19.
Agassiz, Louis. “The Diversity of Origin of the Human Races.” The Christian Examiner and Religious Miscellany 49:110–145. Jul 1850. Agassiz's fullest argument that the human races were created separately, "in nations", in the zoological provinces they now occupy. He affirms a moral unity of mankind while denying common descent, reads Genesis as the history of the white race, and ends with a ranking of races by temperament and a recommendation against treating them as equals.
1. 19a.
  Page 128: “We maintain, that, like all other organized beings, mankind cannot have originated in single individuals, but must have been created in that numeric harmony which is characteristic of each species; men must have originated in nations, as the bees have originated in swarms”. ↩
20.
Darwin, Charles. The Descent of Man, and Selection in Relation to Sex. Volume 1. John Murray, London. 1871. The first edition of Darwin's book on human evolution. Chapter 7 weighs whether the races are separate species and concludes that they are sub-species of one; chapter 5 discusses natural selection in civilized nations.
1. 20a.
  Page 235: “Finally, we may conclude that when the principles of evolution are generally accepted, as they surely will be before long, the dispute between the monogenists and the polygenists will die a silent and unobserved death.” ↩
2. 20b.
  Page 201: “At some future period, not very distant as measured by centuries, the civilised races of man will almost certainly exterminate and replace throughout the world the savage races. At the same time the anthropomorphous apes, as Professor Schaaffhausen has remarked, will no doubt be exterminated. The break will then be rendered wider, for it will intervene between man in a more civilised state, as we may hope, than the Caucasian, and some ape as low as a baboon, instead of as at present between the negro or Australian and the gorilla.” ↩ ↩
21.
Stanton, William. The Leopard's Spots: Scientific Attitudes toward Race in America, 1815–59. University of Chicago Press. 1960. The standard history of the American School of ethnology, from Morton to the eve of the Civil War. Stanton follows the polygenists through their letters and reviews and argues that the South declined their science because it contradicted Scripture.
1. 21a.
  Page 194: “While the North defected, the South had the opportunity to keep faith with the Jeffersonian ideal by grounding the defense of slavery upon a scientific basis. It did nothing of the kind. Southerners discerned in the doctrine of multiple origins an assault upon orthodox religion (a shrewd enough interpretation) and chose to hold fast to the latter. Their devotion to Scripture was not entirely disinterested. They were busily reading about Paul's request for the return of Onesimus, an obvious foreshadowing of the Fugitive Slave Law, and pointing importantly to the curse on Ham in the Book of Genesis.” ↩
22.
Morris, Henry M. The Genesis Record: A Scientific and Devotional Commentary on the Book of Beginnings. Baker Book House, Grand Rapids. 1976. Morris's verse-by-verse commentary on Genesis. Its pages on Genesis 9:24-27 reject the curse as a warrant for slavery and the idea of race, and then assign characters to the descendants of Shem, Ham and Japheth, following Arthur Custance.
1. 22a.
  Page 241: “Sometimes the Hamites, especially the Negroes, have even become actual slaves to the others. Possessed of a genetic character concerned mainly with mundane, practical matters, they have often eventually been displaced by the intellectual and philosophical acumen of the Japhethites and the religious zeal of the Semites.” Morris cites Custance's Noah's Three Sons in a note on the same page. ↩ ↩
23.
Ham, Ken et al. One Blood: The Biblical Answer to Racism. Master Books, Green Forest, Ark. 1999. A popular book by the founders of Answers in Genesis and Creation Ministries International. It argues that all humans are one race descended from Adam, that the curse of Ham is a misreading, and that no biblical rule forbids marriage between people groups.
1. 23a.
  Page 103: “The curse of Canaan has nothing whatsoever to do with skin color”. ↩ ↩
24.
Computed here. Under genetic drift alone, in isolated groups of N breeding adults that start out identical, the fixation index after t generations is F = 1 − (1 − 1/2N)^t, very close to 1 − e^(−t/2N), so N = t ÷ (−2 ln(1 − F)). Barbujani's 1997 study puts 84.4 percent of DNA variation within populations, leaving F = 0.156 between them.⁴ Then −ln(0.844) = 0.170 and N = t ÷ 0.339. For a Babel dispersal 4,000 to 5,000 years ago at 25 to 30 years a generation, t = 130 to 200 and N = 383 to 590. For the mainstream's roughly 2,000 generations, the order of the recombination count since Neanderthal admixture reported in Out of Africa or Out of Babel?, N = 5,896. The calculation assumes constant size, no migration and no selection. Migration between groups lowers F, so with migration the groups would have to be smaller; a founder bottleneck at the split raises F, so with one they could be larger. It treats Barbujani's between-population share of variance as a fixation index, which is the quantity that kind of analysis of variance estimates. ↩ ↩